In this study, a bi-variate Generalized Autoregressive Conditional Heteroscedasticty model is used in order to investigate the Granger causality relationships between output growth, inflation rate and their uncertainties. Our test results show that the existence of Granger-causality is observed from nominal uncertainty to inflation, from nominal uncertainty to real uncertainty, from output growth to real uncertainty, from output growth to nominal uncertainty and from inflation to nominal uncertainty. These findings prove that theoretical predictions of Cuikerman and Meltzer (1986), Okun (1971) and Friedman (1977) are valid for the period 1986:6-2007:1 for Turkey. On the other hand, ‘Short-run Phillips Curve’ and ‘Taylor Effect’ have proven empirically to be invalid for Turkey for this sample period. Moreover, we deduce that Turkish inflation is affected by the output growth through the nominal uncertainty channel.